Article

An Online Retinal Disease Management Module

A module assists ophthalmologists in treating retinal disease by providing clinical recommendations and resources.

The Retinal Assistant Module (www.amretina.tk ) is a web-based tool that uses artificial intelligence to help categorize symptoms and triage cases of diabetic retinopathy and retinal vein occlusion. It offers a global community of ophthalmologists automated management recommendations based on their choice of the clinical scenario via several steps that mimic their cases. Artificial intelligence is programed to accommodate various possible real-life scenarios, including complex cases where previous treatment has failed and combined cases when there is diabetic macular edema with proliferative or nonproliferative diabetic retinopathy, at different levels of severity.

Developing the Module

I was inspired to develop the module because many of my young colleagues were sending their retinal cases to me for online consultation using LinkedIn, Telegram, WhatsApp, and Facebook. Because it was taking so much time to respond, I wanted to create a way to share clinical advice quickly, at any time, online, and for free. I decided to build an online module.

After the idea matured in my mind, I contacted local programmers to build the artificial intelligence system. I could not find a programmer who understood my goal and was able to do what I needed, so I decided to create it myself. I spent weeks learning to code. I programmed the artificial intelligence using JavaScript, CSS, and HTML, and set clinical parameters using information from DRCR.net, the Early Treatment of Diabetic Retinopathy Study, PubMed, and personal experience. After 2 solid weeks of work, the module was built, and I launched it in 2016. The system includes a main module in which an ophthalmologist can choose from a variety of clinical findings to narrow down to a recommendation, which includes a list of related literature and resources. The module also includes an image feature with a library of images that ophthalmologists can compare their patients’ images to (Figure 1), videos, and additional resources.

Figure 1. Within the imaging module, images are presented so that users can compare them to their own findings.

After building the module, I presented it to the International Council of Ophthalmology, which added the module as a resource on its website at http://www.icoph.org/resources/371/Retinal-Assistant-Module.html . The module was also endorsed by Muthusamy Virtual University of Post Graduate Ophthalmology.

To get the word out, I shared the module in ophthalmic groups on Facebook, where it got a good deal of traction. Many colleagues began using it and providing positive feedback. Young ophthalmologists are now using it to help them manage retinal cases by obtaining online recommendations for treatment and follow-up. To date, the Retinal Assistant Module has had about 4,500 visits from ophthalmologists around the world, from countries including the United States, Canada, Russia, Syria, and India.

Updating the Module

A great benefit of this online module is that it has the flexibility to be updated at any time, reflecting the most up to date data and guidelines. I have updated the module several times. I have also tested an Android app so it can be used without the need of a browser. The app is still in development.

Future Modifications

In the future I hope to improve the service by adding diseases like AMD, posterior uveitis, hereditary retinal disorders, and peripheral retinal diseases. I also hope to make it available via the Facebook app, so ophthalmologists can access it without the need of a web browser. Another goal is to create a way for ophthalmologists to upload fundus photographs, fluorescein angiograms, and OCT scans along with case history that the module will analyze.